Tag: long tail
Connect With Us
Customer Buying Patterns – What you can learn From Pizza Sales
There are 1,140 ways of ordering a pizza with 3-toppings, if the pizza offers 20 choices
First order take rates tell us about the relative popularity of different options. For example, consider a small set of possible pizza toppings.
|
Topping |
Take Rate |
|
Pepperoni |
40% |
|
Mushrooms |
20% |
|
Pineapple |
3% |
|
Canadian Bacon |
3% |
|
Green Peppers |
10% |
Customer buying patterns really start with second order take rates, which tell us about pairs of options, or toppings. Second order take rates tell us about relative popularity, but they also reveal something deeper: dependence. If you know that a pizza has pineapple on it, there is a very good chance that it also has Canadian bacon. This is dependence. In this case the reason is that there is a widely known “Hawaiian Pizza” that has both of these toppings. In general, customers don’t flip coins or roll dice. They select options that “hang together” in some way. The patterns can be seen in the combined take rates. Let me illustrate with three examples that are contrived to illustrate some important points. First consider pepperoni and mushrooms together.
|
|
|
Mushrooms |
|
|
|
|
No |
Yes |
|
Pepperoni |
No |
18% |
12% |
|
Yes |
32% |
8% |
|
In this table you can see that pepperoni has a 40% take rate, since 32% of pizzas have pepperoni without mushrooms, and 8% have pepperoni with mushrooms. In the same way, 20% have mushrooms, because 12% have mushrooms without pepperoni and 8% have mushrooms with pepperoni. This illustrates the first law of second order take rates: the first order take rates must be preserved. But this table contains no new information. Customers are apparently ordering pepperoni and mushrooms independently. This is revealed by the fact that 8% is exactly 20% of 40%. Knowing that a pizza has pepperoni does not give us any clue about whether or not it has mushrooms. Similarly, knowing that it has mushrooms is useless in guessing if it has pepperoni.
As a second example, consider the two toppings that are on every Hawaiian pizza: pineapple and Canadian bacon.
|
|
|
Canadian Bacon |
|
|
|
|
No |
Yes |
|
Pineapple |
No |
97% |
0% |
|
|
Yes |
0% |
3% |
For simplicity, I have made this an example of complete dependence: a pizza has pineapple if and only if it has Canadian bacon. Notice that the first order take rates (3% for each) are preserved.
The third example is the really important one: partial dependence. This is illustrated here by pineapple and green peppers.
|
|
|
Green Peppers |
|
|
|
|
No |
Yes |
|
Pineapple |
No |
89% |
8% |
|
Yes |
1% |
2% |
|
In this case, the first order take rates are also preserved: 3% for pineapple and 10% for green peppers. But these two choices are not independent. The take rate for pineapple and green peppers together is 2%, which is much greater than 10% of 3%, which would be only 0.3%.
Exercise for the reader: show that if we know that the pizza has green peppers, then there is a 20% chance that it has pineapple. (Much greater than its 3% first order take rate.) And if we know that it has pineapple, then there is a whopping 67% chance that it has green peppers!
So second order take rates capture information about how customers are combining toppings, and we can use that information to make predictions.
Comments Off
What does the stairway to complexity tell us?
If a product is too complex, where is the complexity coming from? Which features are causing the explosion in the number of build combinations? The stairway to complexity tells us where to look.
The stairway to complexity shows how the number of unique configurations drops as features are removed. Here is another stairway for a backhoe with 30 features.
The number of build combinations drops from 934 down to 1 as we remove the features. Behind the graph is the actual list of features in the order they were removed. In the table below, the features are ranked from 1 to 30, corresponding to the steps in the graph.
If we want to simplify our product, this ranking of the features tells us where to start. The greatest contributor to complexity is the Buckets, of which there are 34 different kinds. The number of build combinations would drop from 934 to 838 if we didn’t have to worry about Buckets.
Is the ordering of the features in the stairway the same as the ordering by number of options? The first feature in the stairway is certainly the one with the most options (34). But Tran_Control has the second largest number of options (9), and doesn’t appear in the stairway until step 15. So there is more going on than just the number of options.
The amount of complexity introduced by a feature depends not just on the number of options, but on the relative popularity of the different options. Having two options that are split 50% to 50% is much worse than if they are split 90% to 10%. (See earlier post: Entropy of a coin toss.)
Introducing a new feature only increases product complexity if it splits existing configurations that would otherwise be the same. One manufacturer insisted that his product was so complex because it was produced for many different countries. But the number of unique build combinations was exactly the same whether the Country feature was included or not.
Comments Off
Self-Service simplifies Product Offerings and increases Margins
Self service is a term we all know, such as pay-at-the-pump gas and self-checkout stations at some grocery stores, and now more obscure things like video game kiosks by GameFly, but the true tidal wave of self-service hasn’t even started, and it’s going to be good for both the consumer and the manufacturer, if done right.
Self Service Grocery Scanner
When you checkout your soda and cereal by swiping products across a scanner at the auto-checkout stations, there isn’t much complexity other than when you get a problem with the scanner reading a smudged bar code or trying to locate the button for ‘snap beans’ when you put those on the scale. The transaction is smooth, quick and you are in control, which is a good feeling as a buyer, you are not being sold, you are buying just what you want, quickly and easily.
But what happens if you try to buy a “configurable product“? In the grocery store, the only thing configurable is the weight of produce, but other than that, the costs and configurations are set in stone and are detected by reading the bar codes. Easy to understand as the buyer and relatively easy to deal with as the seller. Configurable products are those where you have to make many choices before you can order the one product. Products like computers, cars and thousands of others where the buyer has to describe their preferences or choices so the product can be created and delivered. It’s even more complex in a B2B environment than it is in B2C, where the products available and choices are astronomical. Products like Lighting, Valves, Agriculture and Construction Equipment, Lifts, Electrical equipment, cooking equipment and conveyors have more choices and variants than you can imagine and that variety makes it hard to order, build and deliver efficiently.
Usually a large direct sales force is sent out with complex price books (sometimes online in PDF form) to sit with customers and prospects and help them combine choices in hopefully valid ways. The choices a customer have to make are quite extensive, ranging from tens to hundreds of choices. Most of these choices the customer doesn’t care about, but they are required by the manufacturer just so they can build a valid product. Customers care about the few things that matter to them but after that, they will just choose things that “seem to make sense” just to complete the order. Sometimes they don’t even do that, they get so frustrated with 60 more questions about features and options on the product (many of which they don’t understand) that they walk away.
In some cases companies believe that putting in a configurator is the solution to their problem. Configurator’s automate the order process by ensuring that the order is VALID. The engineering and marketing rules that drive what can be built and offered are setup in a configurator such that the user ordering the product is led through valid questions and end up with a build-able product. Now this product may be build-able but it also may be a one-off low-margin brand new SKU that manufacturing hasn’t built before and requires some parts they aren’t carrying at this time. All this for something that was only 2 choices from a very popular configuration. And those 2 differences only happened because the customer was asked 20 more questions after they entered the 5 things they cared about. They chose as best they could, but without any guidance or suggestions, ended up on a new SKU which will ultimately explode into huge numbers of parts and processes to support the new SKU.
Now if the customer only had to enter the 5 things they cared about and the system recommended the combination of other choices such that the customer’s price limit was met and the configuration wasn’t a new SKU and the SKU had a good margin, then it would have been a win-win for everyone. And the whole process could be complete quickly and easily. The customer wouldn’t have to answer any other questions and would feel that same feeling that you do when you swipe your can of soup across the scanner at the market. The manufacturer wins as well because the customer was guided toward an existing configuration so the cost of creating and supporting a new SKU was avoided. It’s happening now with recommendation engines that leverage buying patterns to suggest full configurations based on the few attributes a customer gives it. Just like Amazon can recommend other books you might want to read based on the current “fly fishing” book you are looking at now, suggestion engines can be utilized to provide this convenience for much more complex products.
That’s the self-service tidal wave that’s coming, when all products, not matter how complicated can be ordered by simply asking for the attributes that YOU care about, what your price limit is and then Voila! it’s done. Customers will order more from companies that offer this convenience. Just think about how often you walk into the gas station to pay as opposed to pay at the pump. And if you had two stations to fill up at, one was pay at the pump and the other required you stand in line after pumping the gas, which do you think you would most often go to? Simplification is good for everyone, and profitable too.
Comments Off
How I want to buy a car
Every five or so years, I shop for a new car. I hate car shopping. The haggling, the long trips to dealerships way outside of town, the hours and hours of waiting, punctuated by furtive whispers to my husband, “Don’t give in! Stick to our budget! But don’t tell them our budget!” and similar. But that’s toward the end of the process. There’s a lot of work leading up to it.
First I hit the Consumer Reports site to research cars. A subscription is just $5.95 a month, but it auto-renews so you have to remember to unsubscribe or it quietly chips away at your wallet forever.
I find the five safest vehicles according to my car type and year. When I say new car, I just mean it’s new to me. I like to benefit from someone else’s new-car depreciation, which is something like 25% the minute you drive off the lot.
Anyway, I get on several different car sites like CarsDirect.com and AutoTrader.com to look for my next set of wheels. First I have to pick make and model, then enter my ZIP Code, then there’s a long list of cars. If I want to, I can see the list from lowest price to highest. The trouble is, I want to compare five different models and several different years. I’ve got to select the same filters over and over for all five and then compare the info. continue reading »
Comments Off
Q&A with Mark Gottfredson, Bain & Company
In today’s post, we talk to Mark Gottfredson about product complexity and customer choice.
Emcien: It’s natural for companies to add products and features to keep customers happy. What are the downfalls?
MG: The challenge of adding complexity is it’s the most natural thing in the world. Marketing comes up with new ideas for products or configurations to get the next bit of market share or a little bit more share of wallet. But most companies aren’t so good at retiring products; they don’t have a similarly robust process for taking things out of the catalog that no longer sell, or sell only small amounts. They don’t do a good job of balancing.
Most decisions we make are based on incremental economics. Each decision makes sense in its own right, but the costs of complexity tend to grow systemically. You can’t tie them to a single product decision. Take tinted windshields, for example, that you can sell as an option for $120 and 40% of customers will buy. Assuming the costs of tinting the windshield including inventory impacts, etc., are $9, it will always make sense to add the option. By itself, it is a rational decision, but when coupled with hundreds of other decisions, we end up with dozens of options like power windows, 13 exterior colors, 10 interior colors, 7 different radio and speaker combinations, etc. Eventually, the vehicle can be made in 10 billion different ways, and you don’t know what the next order will be. Since you can’t effectively forecast anymore, you get frustrated and buy a $50 million forecasting module to try to manage all the complexity. You have difficulty balancing your lines, build inventory and increase supply chain costs. Unfortunately, when most companies finally decide to reduce complexity, they “cut off the tail” of low-running options or SKUs. But they don’t remove the systemic costs, and they don’t see any benefits.
Emcien: Companies often overestimate the value buyers place on having many choices. What are the downsides?
MG: Go to a banking website like Citibank or Bank of America. The site describes itself as a full-service bank that has all the items you could want. There are long lists of products like credit cards with different reward programs, as if to say, “We have a lot of products. Surely there’s one here for you. Good luck finding it.” High complexity is a priori evidence that you don’t know what your customers want.
Emcien: When do fewer choices mean higher sales?
MG: When you understand customers. Dell understands customers well. Dell’s website is Spartan; there are just a few choices. If you choose a desktop, up pops three computers: high, medium and low cost. These three configurations are what your segment – home, professional, government – wants. You can customize each one, but you’ll make it as expensive as the next higher model, so then you switch to that and you’re still buying a standard configuration. Every time I have seen complexity reduction done right, sales have increased.
Emcien: How do overoptimistic sales expectations help to spread complexity?
MG: What happens is sales looks for a gimmick that gets them the next sale. Many manufacturers think whatever’s thrown over the wall from product management and sales must be good to go. And sales thinks more is better! Engineers love to engineer; they’ll give you complexity. Most firms build complexity systematically into operations, and then they build systems to handle the complexity, and that’s high cost.
Companies should think about what business would be like with a zero-complexity baseline – how they would operate if they offered just one product or service. The purpose of zero-based thinking isn’t to eradicate complexity; it’s an exercise to reimagine the business with the optimum amount of complexity.
Mark Gottfredson is a director of Bain & Company’s office in Dallas, Texas, which he founded in 1990. Over the past 26 years, he has advised chief executives and top-level managers in a wide range of industries. Currently, he serves as the Global Head of Bain’s Performance Improvement Practice and is also a leader in the firm’s business strategy, airline, financial services, manufacturing and energy practices.
Comments Off
Variation is valuable
Advances in interconnection technologies are driving an increasingly demand-driven market. Customers are learning to expect to get what they want, when they want it, how they want it. And they tell you in each and every interaction they have with your company, or not. In a demand-driven world, increasing product variation and complexity in your business model is inevitable. Left untended, your business can become a tangled web of counterproductive business strategies with a dense portfolio of product families comprising thousands, even millions, of variants.
However, make no mistake, variation is valuable. To deny complexity or view the long tail of product variation as a management failure is to deny diversity of the world in which we make our living. Eliminate complexity in your product offer and you will find yourself competing with boatloads of product from China, India or any of a number of low-wage production markets.
The “keep it simple” principle is the root of good management. However, as Oliver Wendell Holmes, Jr. has observed, “I would not give a fig for the simplicity this side of complexity, but I would give my life for the simplicity on the other side of complexity,” it matters which form of simplicity you choose. The wrong simple answer is to try to focus on the 20% of product variants that make up 80% of your revenue, the head of the ubiquitous Pareto distribution, and find ways to minimize or eliminate the so-called unprofitable remaining 80% of product variants that lurk in the tail. Hello commodity, goodbye margins. The right simple answer is to deliver Intelligent Variation based on the voice of the customer shouting through the many interactions they have with you each and every day.
Comments Off
The typical tail graph
In a previous post, I discussed two types of sales history: raw and collapsed. The collapsed sales history can be displayed in a table or spreadsheet, with a special column for volume. If this table is sorted on decreasing volume, then the most popular configurations (popcons) will be at the top. The graph with the volumes displayed in decreasing order (popcons on the left) is called the tail graph
We have drawn tail graphs for cars, computers, washing machines, lighting fixtures, trucks and tractors, and they all look basically the same. The first tail graph shown below is small but typical. It represents 2,884 tractors, with 1,997 unique configurations, or build combinations. On average, there are 1.44 units per unique configuration. The most popular configuration was ordered 23 times. The graph quickly drops to two of a kind and finally one of a kind (our technical terms are “twosies” and “onesies”). Combined, the onesies and twosies account for 2,000 tractors, or 69% of total volume. Rather high, though this number is usually at least 40%.
Comments Off
Understand product choices to manage complexity
A product is a collection of features, and each feature has alternative options. Understanding features helps determine strategies such as late staging. Some features are tangible, material things about the product: which engine, how much memory, Bluetooth. Other features are abstract or soft, like geographic region or sales channel. Among tangible features, distinctions can be made on the basis of the degree of postponement possible.
Pin-on features can be added to the product at the last minute, after a specific order is received. The classic example is the power cord for a printer. Hewlett-Packard avoided having different printers for different countries by attaching different power cords to a common printer.
Reconfigurable features can be changed after a real customer order is received. There are literally hundreds of different kinds of tractor tires, depending on the work a tractor will be used for. A tractor has to be built with some kind of tire just so it can be driven off the assembly line, and it’s easy to change the tires to suit the customer.
Line features, by contrast, are so basic to the product that they can’t be changed, such as the chassis or transmission for a vehicle or the motherboard for a computer.
Abstract or soft features are really attributes of the order rather than of the product itself. But they may be very valuable in understanding customer demand. The pattern of choices for tangible features may vary considerably by geographic region, which is a soft feature. For example, engine block heaters are popular in North Dakota and convertibles are popular in Florida, but convertibles with engine block heaters are almost non-existent.
Comments Off
Extending the product configuration to gain insight
One of the most important components in choice complexity is the product configuration itself, the mixture of product options that give a product its unique signature. Obviously the typical product orderable options are needed to analyze the complexity of a product, but other more abstract options can offer surprising insights into product and customer behaviors.
A typical car configuration has options such as sedan, V6 engine, automatic, blue, cloth, AM/FM/CD, sunroof. But more abstract items can be recorded along with these to offer more insight. Sales type can be recorded to analyze what types of product configurations sell better in promotional sales events as opposed to normal sales transactions. An attribute to record an extended factory warranty option may provide new ideas for packaging options together with additional warranty services that customers are moving towards.
Comments Off
Stop product complexity at the door
In any manufacturing company that builds configurable products, there is a lot of discussion around what product complexity is. What’s interesting is that when times are good and there are lots of sales, the discussion is usually around how to simplify or streamline with the goal to sell more product even faster, that complexity is keeping sales from going even higher. In bad times, the discussion typically moves to how complexity is causing undue stress on the supply chain, creating problems with parts forecasting, quality and finished goods inventory.
Rarely do these discussions end with participants really agreeing about exactly what complexity is or how to reduce it. Solutions are attempted with internal projects like SKU reduction and part number reduction initiatives driven by Six Sigma teams that mean well and do good work, but usually are chasing the tail of the complexity dog, rather than leashing it for good and guiding it to higher profits, lower forecasting errors, even shorter sales cycles.
Comments Off